|
Items in cart [0] |
Questions? Call 888-624-8373 |
| Copyright © 2009. National Academy of Sciences. All rights reserved. Terms of Use and Privacy Statement |
Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter.
Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 17
Motor Vehicle
Emissions: A Strategy
for Quantifying Risk
RICHARD R. BATES
ANN Y. WATSON
Health Effects Institute
Automotive Emissions: A Brief Perspective / 19
Understanding the Components of Risk / 19
Chemical and Physical Properties of Pollutants / 20 Pollutant
Exposure and Dose to Biological Tissues / 22 Biological Responses
to Pollutants / 23 Sensitive Individuals / 24
Research Approaches / 24
Epidemiology / 24 Experimental Methods / 25 Mathematical
Models / 26 Physical Models / 26
Pursuing Quantitative Toxicology: A Strategy for Research / 26
Highlights of the Authors' Recommendations / 27 Criteria for
Prioritization / 27 A Proposed Research Strategy / 28
Summary 1 34
Conclusion / 34
Air Pollution, the Automobile, and Public Health. @) 1988 by the Health Effects
Institute. National Academy Press, Washington, D.C.
17
OCR for page 18
18
A Strategy for Quantifying Risk
The ultimate objective of a regulatory-
oriented research program that focuses on
air pollution from mobile sources is to
identify and quantify any effects that those
emissions may have on human health. But
before we invest intellectual and financial
resources, we must first understand the
limitations of current information and
methodologies that preclude accurate esti-
mates of risk to human health. Future
research programs should be justified by
their promise to overcome these limita-
tions. The goal of this volume, then, is to
identify issues and select a research agenda
that will be most effective in advancing our
ability to quantify the health risks associ-
ated with air pollution.
To understand risk, it is essential to
understand the relation between the
amount or dose of pollutants, and the
response, in terms of human illness, to
those pollutants (National Research Coun-
cil 1983~. Yet the obstacles that challenge
our understanding of that relation are for-
midable. Biologists ask, "To what sub-
stances are people exposed, and what dose
levels are most relevant for experimental
studies?" Engineers and physical chemists
inquire, "Which of the hundreds of com-
pounds and chemical reactions are biologi-
cally important?" This volume alone cannot
answer these questions, but its discussions
are critical to beginning a dialogue among the
. . . , . .
various sclentl~lc commumtles.
The first part of this book addresses
research about the exposure of humans to
vehicular emissions and their reaction
products. The first three chapters examine
the nature of automotive emissions John-
son), their chemical transformation (At-
kinson), and the mechanisms of their trans-
port through the atmosphere (Samson).
Subsequent chapters explore the use of
ambient measurements (Graedel) and math-
ematical models (Russell) as ways of de-
scribing pollutant concentrations, which in
turn are key descriptors of human exposure
(Sexton and Ryan). The remaining chapters
discuss the principles and mechanisms by
which, during exposure, particulate (Schle-
singer; Sun, Bond, and Dahl) or gaseous
(Overtop and Miller; Ultman) pollutants
are inhaled and distributed throughout the
respiratory tract.
In the second half of this book, the
chapters examine whether specific biologi-
cal responses to airborne pollutants consti-
tute a health hazard. The problem is ap-
proached trom two complementary points
of view. One view starts with the chemical
constituents of automotive emissions and
their transformation products and investi-
gates their effects on human health. In this
approach, oxidants (Bresnitz and Rest),
polycyclic aromatic hydrocarbons (PAHs)
and nitro-PAHs (Hecht), and alcohols and
aldehydes (Marnett) are discussed.
In contrast, the second view begins with
a set of human diseases and investigates
whether automotive emissions or their de-
rivatives play a role in disease develop-
ment. Attention is given to health effects of
obvious relevance to airborne contaminant
exposure. Thus, asthma (Bromberg), res-
piratory infection (Pennington), cancer
(Kaufman), fibrosis (Last), and emphysema
and small airways disease (Wright) are con-
sidered. Coronary heart disease (Clarkson)
is examined because cigarette smoke in-
creases mortality, and it is not known if
other combustion products affect morbidity
and mortality rates. Neurobehavioral effects
(Wood) of hazardous substances, which may
not be an obvious health effect, also warrant
investigation. One's perception of, and re-
sponse to, the environment may be just as
significant for the "health" of society as clin-
ical presentations of organic disease.
Thus, this book examines the existing
evidence about risks from automotive
emissions and evaluates the methodologies
presently available to quantify human risks.
However, not all possibilities are explored.
Areas for which the evidence was lacking
or the rationale for associating automotive
emission exposure with health risks was
not compelling were omitted; examples
include the role of automotive-derived pol-
lutants in the etiology of reproductive dis-
orders or renal dysfunction. By the same
token, we did not want to duplicate previ-
ous analyses; for example, numerous pol-
lutants are described in the U.S. Environ-
mental Protection Agency (EPA) Criteria
Documents. Unfortunately, even though
human exposure is not to a single material
or even to a uniform mixture, we were
unable to properly explore the issue of
OCR for page 19
Richard R. Bates and Ann Y. Watson
19
complex mixtures. Many of the chapters
stress the necessity of acquiring better in-
formation about atmospheric constituents
and quantitative markers of human disor-
ders before a research protocol can be for-
mulated to test, within the scope of limited
resources, the bioactivity of such mixtures.
Although the primary focus of this book
is automotive emissions a focus fully jus-
tified by the omnipresence of motor vehi-
cles in our society the scientific principles
underlying many of the discussions are
applicable to other air pollutants and even
to other sources of exposure. First, auto
emissions and their derivatives are not the
only source, or even the predominant
-
~ ized.
source, ot ~ auman exposure to many ot
these chemical substances. Second, air pol-
lutants, whether they originate from auto-
mobiles, power plants, or wood-burning
stoves, are all governed by similiar princi-
ples of transport through the atmosphere
and into the body. Finally, the research
needed to identify people who are suscep-
tible to emphysema, pulmonary fibrosis, or
lung cancer is similar whether the causative
agents derive from auto emissions, occupa-
tional exposures, or cigarettes.
The primary purpose of this overview
chapter is threefold: (1) to describe the
issues and difficulties involved in providing
reliable scientific data for risk assessment;
(2) to highlight and integrate the authors'
recommendations into the issues raised;
and (3) to propose a research strategy. We
start with putting into perspective the au-
tomobile as a source of potentially harmful
substances. We next describe components
that determine risk of injury and discuss the
strengths and weaknesses of various research
approaches. We then summarize the authors'
choices of future research directions. Finally,
after having carefully read the authors' sug-
gestions, we conclude with an attempt to
translate their individual priorities into a fea-
sible, cross-disciplinary research strategy.
Automotive Emissions:
a Brief Perspective
Motor vehicles contribute significantly to
. . . . . . . ,-
emlsslon Inventories In certain regions ot
the United States, particularly in urban
areas. Table 1 shows the relative contribu-
tion of various emission sources to the total
inventory of five pollutants for the country
as a whole and for two metropolitan areas.
Typically, in urban areas, motor vehicle
emissions dominate carbon monoxide in-
ventories and can contribute up to 50 per-
cent or more of nitrogen oxides and hydro-
carbons. It is, however, important to
remember that these figures reflect outdoor
pollutant concentrations. Most people in
industrialized societies spend more time
indoors than outdoors, and the contribu-
tion of motor vehicles to pollution of in-
door air has not been adequately character
l wo additional points are worth noting.
First, volatile organic compounds and total
suspended particulates are broad categories
that include many individual compounds of
divergent toxicologic importance. Thus,
table 1 may not adequately portray the
relative impact of these emission sources on
human health. Second, comparable data are
not available for unregulated emissions
which include many compounds of possi-
ble toxicologic interest. Therefore, even
though motor vehicles emit numerous
chemical species into the environment, we
cannot determine, with current data bases,
the relative contribution of mobile sources
to the health risks of our society.
Understanding the
Components of Risk
Pollutants and their derivatives cause harm
by interacting with, and impairing, mol-
ecules crucial to the biochemical or phys-
iological processes of the human body.
Figure 1 illustrates the pathways from pol-
lutant sources to toxic effects. Clearly, it is
difficult to link a specific pollutant to a
particular ill effect until the processes in
between are understood. The risk of toxic
injury from a substance depends on three
factors: the chemical and physical proper-
ties of the substance, the dose of the mate-
rial that reaches critical tissue sites, and the
responsiveness of these biological sites to
the substance. The relationship of each
OCR for page 20
20
A Strategy for Quantifying Risk
Table 1. Relative Contribution of Various Sources to Emissions of Regulated Pollutants
United States: 1984a
106 tons/year (%)
Volatile Total
Carbon Nitrogen Organic Suspended Sulfur
Source Monoxide Oxides Compounds Particulates Oxides
Transportationb 48.5 (69) 8 7 (44) 7 2 (33) 1.3 (19) 0.9 (4)
Fuel combustionC 8 3 (12) 10.1 (51) 2 6 (12) 2 0 (29) 17.3 (81)
Industrial process 4.9 (7) 0.6 (3) 8 4 (39) 2 5 (36) 3 1 (15)
Solid waste disposal 1.9 (3) 0.1 (1) 0.6 (3) 0.3 (4) 0.0 (0)
Other 6.3 (9) 0.2 (1) 2 7 (13) 0.9 (13) 0 0 (0)
Metropolitan Los Angeles: 1981
103 tons/year (%)
Transportationb 3,859 (92) 399 (62) 406 (44) 431 (59) 57 (26)
Fuel combustionC 45 (1) 193 (30) 27 (3) 20 (3) 64 (29)
Industrial process 110 (3) 48 (7) . 223 (24) 31 (4) 97 (44)
Solid waste disposal 116 (3) 2 (I) 39 (4) 16 (2) 0 (0)
Other 76 (2) 1 (1) 239 (26) 235 (32) 0 (0)
New Jersey-New York-Connecticut: 1981
103 tons/year (%)
Transportationb 4,248 (94) 349 (58) 397 (38) 361 (53) 50 (12)
Fuel combustionC 91 (2) 229 (38) 26 (2) 42 (6) 315 (75)
Industrial process 26 (1) 9 (2) 276 (27) 194 (29) 47 (11)
Solid waste disposal 125 (3) 10 (2) 31 (3) 41 (6) 7 (2)
Other 24 (1) 1 (1) 311 (30) 39 (6) 0 (0)
a From U.S. Environmental Protection Agency (1986).
b Includes motor vehicles, rail, aircraft, vessels.
c Includes stationary sources such as residential, electric generation, industrial, commercial-institutional.
~ From U.S. Environmental Protection Agency (1984).
component to the scheme presented in fig-
ure 1 is discussed below.
Chemical and Physical Properties
of Pollutants
Exhaust from the tailpipe of motor vehicles
is a complex mixture that contains hun-
dreds of chemicals in the form of gases as
well as solid and liquid aerosols. The com-
position of the mixture depends on the fuel,
the type and operating conditions of the
engine, and the effects of any emission
control devices. Upon their release, emit-
ted substances are transformed by complex
atmospheric chemical reactions. Airborne
pollutants, therefore, consist of primary
tailpipe emissions (for example, carbon
monoxide, nitric oxide), and new chemical
species formed as a result of atmospheric
reactions (for example, nitrogen dioxide,
ozone); the opportunity for chemical diver-
sity is immense. In addition, the chemical
composition is dynamic; the air we breathe
today is different than the air we breathed
10 years ago. Therefore, a changing con-
stituency must be recognized in any evalu-
ation of chronic health effects.
The physical form of airborne contami-
nants will influence their distribution both
in the atmosphere and in biological tissues.
As emissions cool, some vapors are ad-
sorbed onto particles or condensed into
droplets, and fine particles and droplets
coalesce into large ones. Upon inhalation,
the regional tissue distribution of gases or
aerosols depends on such physical proper-
ties as size and solubility. In addition, the
deposition, metabolism, and clearance of
volatile organic compounds are dramati-
cally altered if they are adsorbed onto
respirable particles. Eventual dose, there
OCR for page 21
Richard R. Bates and Ann Y. Watson
21
Fuel
Engine
characteristics
Emission
controls
Other
sources
Mobile
source
emlsslons, ,
Sinks
r~
GENETIC AND
I ENVIRONMENTAL
I MODIFIERS
1
Detoxification
Primary
pollutants
Atmo. Ph. uric
r r
;
I Ind oor
| pollutants |
)
I
Dose to
pulmonary tissues
l
Metabolism
Dose to
/ \ I target tissues
| Cellular and
| molecular cascade l
l
| toxic agents 1 t: -
Figure 1. Pathways from pollutant sources to toxic effects.
fore, is intimately associated with the phys-
ical properties of the airborne contami-
nants.
Because chemical structure and physical
characteristics are important determinants
of toxicity, an improved understanding of
these properties is essential. For example,
slight modifications of functional groups
on PAHs can markedly alter the mutagenic
potential. Nasal, but not bronchial, tumors
have been found in rats after prolonged
exposure to formaldehyde, which, because
of its solubility, is absorbed in the nasal
cavity. In addition, some compounds have
OCR for page 22
22
A Strategy for Quantifying Risk
similar properties but different biological
effects. Ozone and nitrogen dioxide are
oxidants, but only nitrogen dioxide pro-
duces emphysema in laboratory animals.
Other as-yet-unidentified factors must in-
fluence outcome. Information about struc-
ture and toxicity is available for some
classes of closely related materials that have
been studied extensively. For most sub-
stances in the environment, however, data
necessary for quantitative prediction of
toxicity based on chemical or physical
. . .
properties are Sac Equate or unavailable.
Pollutant Exposure and Dose to
Biological Tissues
Pollutants derived from vehicular exhaust
are transported away from their sites of
release by wind and diffusion. Once air-
borne, they mix both with pollutants from
other sources and with materials of natural
origin, and they may be chemically trans-
formed into other species. Ultimately, at-
mospheric pollutants are removed from the
air either by rain or by dry deposition to the
earth's surface. Highly reactive chemicals
may be transformed or removed within a
few minutes; stable substances may persist
for years. Meteorological conditions and
physical structures can also profoundly in-
fluence atmospheric concentrations. Thus,
individual exposure is determined by the
location a person occupies in relation to
emission sources, as well as by patterns of
atmospheric transport, transformation, and
dilution.
Airborne pollutants that are inhaled may
deposit onto surfaces of the respiratory
tract. Deposited insoluble material is moved
(either intra- or extracellularly) toward the
pharynx by mucociliary action and then
swallowed. Alternatively, particles may be
sequestered for long periods within pulmo-
nary tissue or in adjacent lymph nodes.
Inhaled chemicals that dissolve in body
fluids may pass from the respiratory tract
into the bloodstream and circulate through-
out the body. As a result, air pollutants
may affect extrapulmonary organs.
Pollutants may be chemically trans-
formed within the body by metabolic en-
zymes. The liver is particularly active in the
metabolism of foreign chemicals, but the
lung and other organs also have this capac-
ity. In general, metabolism facilitates ex-
cretion of pollutants from the body and
thus reduces pollutant levels in body tis-
sues. In addition the toxic potential of some
parent compounds may be reduced by met-
abolic conversion. Paradoxically, though,
metabolism may also generate products
with increased toxicity. The balance be-
tween metabolic processes that increase
toxicity, decrease toxicity, or favor elimi-
nation is an important factor in the sensi-
tivity of an individual to a toxic chemical.
The definition of "dose" may vary
widely. It may be based on the concentra-
tions of inhaled pollutants at any point
from their deposition on respiratory tract
surfaces, to the concentrations of reactive
material at "target sites," where damage
occurs. The latter is the most valuable
definition of dose and the most difficult to
determine. Dose-to-target measurements
require identification of the sites of dam-
age, understanding the mechanisms by
which the toxic material produces damage,
and knowledge of the reactive materials
responsible for toxicity. Obtaining such
information from laboratory animals or
tissue cultures is not trivial and is seldom
feasible in living human beings.
Instead, for most practical purposes, we
must rely on some surrogate measurement
of dose-to-target sites. This may be the
dose to some other, but more accessible
body tissues, such as the blood or respira-
tory tract. Alternatively, it may be a mea-
sure of exposure, or even of atmospheric
concentration. The more removed the sur-
rogate measurement is from the target site
in humans, the more the approximation is
confounded by intervening environmental
and biological variables. Nevertheless, sur-
rogate measurements may be fully ade-
quate for some applications. In other cases,
mathematical models of the relationships
between exposure and dose can extend the
applicability of surrogate measures. Do-
simetry models that correlate dose between
laboratory animals and humans also broaden
the use of surrogate measures.
Surrogate measures of dose are com-
monly used in experimental and epidemio
OCR for page 23
Richard R. Bates and Ann Y. Watson
23
logic studies. Dose/response studies of tox-
icity in laboratory animals usually measure
exposure, not dose. But differences in the
respiratory tract anatomy and ventilation of
humans and rodents mean that identical
exposure may not result in identical dose.
Such studies, however, may nevertheless
become increasingly important in predict-
ing human health effects as mathematical
models capable of supporting interspecies
extrapolations are developed; these models
aim to relate, in rodents and humans, the
physical and chemical properties of pollu-
tants to the sites and amounts of pollutant
deposited. Interspecies comparisons of ac-
tive metabolite dose or extrapulmonary
dose-to-target sites will require additional
studies of the species-specific pharmacoki-
netics (absorption, metabolism, distribu-
tion, and excretion) of the chemical.
Epidemiologic investigations of the
health effects of air pollutants often exam-
ine the relationships between the health of
people in a community and the concentra-
tions of pollutants at a few stationary mon-
itoring sites within the community. The
people studied are not congregated around
the monitoring stations but are widely sit-
uated, indoors as well as outdoors. A1-
though more localized measurements of
subjects' exposure levels would provide
better data, such measurements are expen-
sive and, for some pollutants, not techni-
cally feasible at this time. Mathematical
models of the dispersion and transforma-
tion of pollutants from their sources, cou-
pled with a better understanding of time/
. .
activity patterns wit nin the community,
would provide an alternative approach for
making better exposure estimates. Im-
proved accuracy in exposure data for the
study population would aid in the applica-
tion of the conclusions to other popula-
tions.
Another area of considerable interest is
the development of indices that reflect the
actual amount of pollutants that reach tar-
get tissues. Although it may not be tech-
nically feasible to measure levels of the
original substance, quantification of metab-
olites or reaction products may be possible.
The discovery that carcinogenic metabo-
lites covalently bind to macromolecules
such as DNA or protein, which in turn can
be measured, has stimulated extensive re-
search in this area. It should be noted,
however, that the use of such approaches
for vehicular emissions is complicated by
the fact that many of the pollutants of
interest are present in very low concentra-
tions. Therefore, more sensitive analytical
techniques for measuring molecular dosim-
eters are needed than are currently being
used.
Biological Responses to Pollutants
The interaction of pollutants with biologi-
cal molecules (often called receptors) trig-
gers the mechanisms of toxic response.
Some responses are direct, the immediate
result of the effect on the receptor. For
example, some forms of asthma may be
precipitated by the direct stimulation of
airway irritant receptors by an inhaled pol-
lutant. In contrast, some other manifesta-
tions of toxicity are highly indirect and,
hence, poorly understood; this is frequently
true of chronic or delayed effects of toxic-
ity. In these cases, the initial interaction of
the toxic chemical with a target site recep-
tor may trigger a cascade of molecular and
cellular events that ultimately damage tis-
sue other than the original target site. Pul-
monary fibrosis and emphysema illustrate
such indirect manifestations of toxicity.
Damage to the pulmonary connective tis-
sue, the hallmark of these diseases, does not
result from the direct actions of inhaled
toxic pollutants, but rather, it results from
a complex, multistep process: pollutant ex-
posure initiates an inflammatory response,
which in turn stimulates excessive produc-
tion of connective tissue or causes the en-
zymatic digestion of elastic tissue. Through
a self-enhancing process, damaged tissue
and inflammatory cells release chemical
mediators that stimulate the recruitment
and proliferation of more inflammatory
cells. Products of these cells include highly
reactive and toxic oxygen-derived radicals,
digestive enzymes, and growth factors.
In addition to the complexity of cellular
and molecular events, it is not clear what
regulates the balance between normal de-
fense or repair functions and abnormal pro
OCR for page 24
24
A Strategy for Quantifying Risk
cesses. Components of the inflammatory
system, which include macrophages, neu-
trophils, and complement, are essential fac-
tors in the defense against infectious agents
and inhaled dusts. These components, in
proper balance under biological control
mechanisms, kill invading microorganisms
and help to prevent or repair tissue damage.
Yet under certain pathological circum-
stances, they produce pulmonary diseases.
Two inferences can be drawn from such
complex and poorly understood mecha-
nisms of toxicity. First, a research strategy
useful for evaluating exposures that is,
tracing the path of a pollutant from its
source to sites of human exposure, and vice
versa-cannot be used for many biological
problems. We simply know too little about
the "middle phase" of most chronic dis-
eases to follow the molecular path from
pollutant deposition to aberrant tissue
structure or function. The situation is fur-
ther complicated because a single pollutant
may trigger a variety of immediate bio-
chemical and physiological actions as it
passes through the body. Many, if not
most, are inconsequential, but it is impos-
sible to recognize the few of toxicologic
importance without understanding the
chronic disease process.
The second implication of the preceding
discussion is that a combination of factors,
including genetics, nutrition, and other en-
vironmental chemicals, influence each of
the multiple stages in the development of
adverse effects, as well as the interactions of
enhancing or inhibiting agents at each
stage. These factors undoubtedly deter-
mine the sensitivity of any person to the
toxic effects of a pollutant. The definition
and enumeration of sensitive individuals
could be pursued more effectively if the
steps between the initial interaction of toxic
substances with their target sites and the
ultimate manifestations of chronic toxicity
were better understood.
Sensitive Individuals
The ability to identify population sub-
groups particularly vulnerable to health ef-
fects induced or exacerbated by hazardous
substances is a critical aspect in the assess-
ment of human exposure to automotive
emissions. In particular, the young or the
elderly may be especially susceptible to
deleterious effects; persons with asthma
may experience aggravated symptoms
upon exposure.
We currently cannot determine whether
differential responses in sensitive individu-
als result from a greater delivered dose or
from inherent biological differences. Im-
proved measurements of exposure and dose
will aid in the identification of people who
exhibit greater responses primarily because
they receive more of the pollutant. The
challenge, however, is to identify those in-
dividuals who are inherently more prone to
disease states. Most likely, the genetic
background of individuals contributes sig-
nificantly to their biological response. The
multiple cellular and biochemical processes
. . . .
set into motion in response to reactive
materials are under complex genetic con-
trol. Taken together, such processes deter-
mine individual sensitivity and, hence, the
outcome from exposure to toxic sub-
stances.
Research Approaches
The utilization of multiple approaches will
most likely enable us to make reasonable
predictions about the risk of toxic injury
from automotive emissions or their constit-
uents. An effective research strategy will
consist of epidemiologic studies, experi-
mental studies, and/or model systems. The
advantages and limitations of these ap-
proaches are described below.
Epidemiology
Epidemiology represents the most direct
approach to demonstrate in humans that
exposure to a particular substance results in
an increase in the incidence of a specific
disease. Limitations in the methodologies
that assess exposures and biological effects,
however, prevent better use of epidemio-
logical approaches. Although epidemio
OCR for page 25
Richard R. Bates and Ann Y. Watson
25
logic studies are valuable for detecting ex-
acerbations of existing chronic disease or
the occurrence of acute respiratory disease,
their use for revealing causes of chronic
respiratory disease is limited. The time lag
between exposure to automotive emission
products and the appearance of a chronic
disease makes it difficult to prove a causal
relationship unless early markers of the
disease are known. It is highly desirable,
therefore, to determine that adducts or other
proposed molecular dosimeters are indeed
reliable indices of exposure and to identify
early markers of chronic disease. People who
should be examined in epidemiologic studies
include those with well-defined exposures,
specific diseases, or inherent factors of sus-
ceptibility. Selection of high-risk groups pro-
vides a more sensitive basis for the detection
of any increase in disease.
Experimental Methods
Clinical. Two important advantages of
human clinical studies are controlled expo-
sure and data derived from human subjects.
The use of exposure chambers permits the
delivery of known quantities of pollutants.
Construction of dose/response curves de-
mands, though, that careful attention be
given to pollutant generation and measure-
ment. Numerous methods that cover a
range of human response, from physiolog-
ical to biochemical effects, are available.
Traditionally, pulmonary function tests
have been used to assess functional capabil-
ities of the lung. Some of these tests form
the clinical definition of disease, such as
asthma or chronic obstructive pulmonary
disease. Because pulmonary function tests
reflect end points of respiratory malfunc-
tion, they have limited value in detecting
the onset of chronic disease. A more recent
innovation, the fiber-optic bronchoscope,
permits access to human tissues by biopsy
and to human cells by ravage. Sampling of Cell Culture. Cell cultures can be used for
human tissues will be of central importance ~r . ~ at a.
to confirm animal studies. Although bron
choscopy can be performed with minimal
risk, it is not ideal for large-scale screening;
therefore, the development of noninvasive
techniques should also be pursued.
Animal. Several issues must be consid-
ered for proper implementation of labora-
tory investigations. Animal studies should
address three components of experimental
design: exposure conditions, disease state,
and extrapolation to humans. Ideally, ani-
mal exposure should mimic conditions of
human exposure. Pollutant concentration
should be low and administered by inhala-
tion. To establish links between acute and
chronic effects, wide ranges of exposure
. . . . . .
regimes are necessary. Once It IS not always
feasible or practical to conduct chronic,
low-dose, inhalation studies, deviations
from these parameters should be validated.
Sedation of laboratory animals should be
avoided, since anesthesia will alter their
ventilatory parameters. An improved un-
derstanding of the influence of chemical
structure on toxicity is necessary to help
guide the choice of appropriate mixtures of
compounds that should be tested.
Experimentally induced disease should
be similar to the human counterpart. In
some cases, it may be necessary to develop
more appropriate animal models for some
forms of toxicity, as well as for susceptible
populations. In addition, if relevant cofac-
tors exist in human ailments, these must be
considered. For example, research on the
effects of air pollutants on morbidity or
mortality from atherosclerosis cannot ig-
nore the role of diet. Likewise, studies on
the susceptibility to infection should use
infectious organisms that are appropriate
models of human pathogens.
Finally, before accurate dose/response
relationships in animal studies can be ap-
plied to human risk evaluation, it will be
necessary to develop reliable extrapolation
or scaling factors. This will require more
extensive baseline information on the rele-
vant anatomy, histology, physiology, and
metabolism in the various species.
~7
the Isolation ot metabolltes or mediators, as
well as for the elucidation of biochemical
mechanisms of toxicity. The use of culture
systems allows compounds or complex
mixtures to be quickly screened for toxic
effects. It is essential.
however, that results
OCR for page 26
26
A Strategy for Quantifying Risk
from in vitro studies be related to in viva
outcomes; for example, the generation of
DNA adducts in culture must ultimately be
correlated with tumor formation in the
whole organism.
Mathematical Models
Mathematical models can be used to esti-
mate or predict emission levels, air quality,
human exposure, and respiratory tract dis-
tribution of inhaled particles and gases. As
alluded to above, the primary asset of these
mathematical models is their potential to
provide a surrogate measure of dose. In
addition, models can be used to distinguish
individual contributions from various
sources to the total pollutant load. Several
models have been developed for these pur-
poses. Such models, though, are only as
good as the chemical and physical data
upon which they are based. There is a need
to improve data bases and to perform
model validation studies. New models also
need to be developed, as well as incorpo-
rating additional concepts into current sys-
tems; of primary importance is the inclu-
sion of indoor air quality into exposure
assessment and source apportionment
models. In addition, predictions of noncri-
teria pollutant concentrations could be im-
proved if better mathematical descriptions
of atmospheric chemical reactions were
used. With respect to dosimetry models,
reactive gas uptake in the upper respiratory
tract and particle deposition In regional
areas are inadequately developed.
Physical Models
Some research questions are poorly suited
to analysis by mathematical modeling
methods. In such cases, it is often easier to
construct a small-scale physical analogy to
the problem of interest, and then test that
scale model to learn how the system be-
haves. Physical models of atmospheric fluid
mechanics problems, such as the dilution of
pollutants near buildings, can be built and
tested in wind tunnels. Atmospheric chem-
ical reactions can be studied in smog cham-
bers that are constructed to replicate the
irradiation of pollutant mixtures by the
sun. Physical models of the geometry of
the human respiratory tract can be used to
study the patterns of aerosol deposition or
gas transport in the lung.
Pursuing Quantitative
Toxicology: A Strategy
for Research
Armed with the knowledge of what is
known about the components of risk, and
of what research approaches are available,
we can focus our attention on how to go
about supplying the missing pieces be-
tween the point of pollutant emission and
the generation of toxic response. Returning
to figure 1, we see that the investigator can
link chemical substances to toxic biological
responses by starting at either end of the
puzzle. In other words, the hypothesis can
be posed in either of two ways: (1) Can
automotive emissions adversely affect hu-
man health under conditions of human
exposure? (2) Do human ailments or phys-
iological malfunctions result from exposure
. . · ~
to automotive emlss1ons'
Numerous factors complicate the design
and interpretation of any experiment. For
example, pursuit of the first hypothesis can
be confounded by the transformation of
primary pollutants and the generation of
new chemical species; the "molecule of
interest" and its concentration are not al-
ways readily identifiable. With respect to
the second hypothesis, the contribution of
other agents capable of inducing disease
makes the assignment of attributable risk to
automotive emissions more difficult. Such
factors cannot be ignored, and any research
strategy must take their presence into ac-
count.
Regardless of the number of chemicals
that might be tested for toxicity or the
number of toxic effects that might be
sought, the evaluation of health effects
lacks a satisfying sense of unity. No satis-
factory paradigm provides a coherent ex-
planation of the relationships between
chemical structure, dose, and toxic effects.
Perhaps such a paradigm could be said to
have existed when the principal concern of
OCR for page 27
Richard R. Bates and Ann Y. Watson
27
toxicology was with immediate and readily
observable toxic actions. For these, the
maxim of the sixteenth century Swiss phy-
sician and alchemist, Paracelsus, applied:
"In all substances there is a poison, and
there is nothing without a poison. It de-
pends only upon the dose whether a poison
is poison or not" (Paracelsus 1958~. Ac-
cording to this guidance, chemicals could
be categorized into groups on the basis of
the magnitude of dose required to cause
death or some other toxic end point (Klaas-
sen and Doull 1980~. Although this precept
served a useful purpose, it is insufficient for
evaluating health effects of automotive
emissions and their derivatives. For health
effects that occur with low frequency or
low magnitude, which are the more likely
result of these materials, the poison is not
only in the dose but also in the genetics,
environment, and lifestyle of the exposed
individual.
An unfortunate consequence of the lack
of a unifying theoretical framework for
toxicology is the phenomenological ap-
proach used for investigating health risks
of pollutants. The battery of tests that
should be done to assess each important
type of health hazard is large and still
growing. The task is magnified immensely
when the pollutant is as complex and as
variable in its composition as automotive
emissions and their transformation prod-
ucts. It is impractical to study the effects of
each chemical individually; nor could such
an approach adequately assess the synergis-
tic and antagonistic influences of the mix-
ture on the toxicity of each component. On
the other hand, it is equally impractical to
test all possible variants of the complex
mixture that would be produced under
differing conditions of fuel, engine opera-
tion, and climate. Moreover, no single set
of conditions can be assumed that might
result in a representative exhaust emission
mixture, which, in turn, could serve as a
toxicity testing standard for experimental
work.
This dilemma will most likely exist for
some time into the future. In the short
term, we can look toward the scientific
developments that provide guidance on the
most important applications of phenome
nological toxicology. Perhaps the develop-
ment of batteries of inexpensive, short-
term tests for the most important toxic
hazards will direct attention to a small
number of the components most in need of
thorough evaluation. Alternatively, such
tests might be applied to varied samples of
the complete mixture to define subsets with
distinctly different toxicological character-
istics. Each subset might then be more fully
investigated.
It is possible, however, to speculate on
directions of research that may eventually
lead to improved quantification of human
exposure and risk. Well-defined and well-
focused investigation can provide some of
the necessary information. But the puzzle
will not be solved tomorrow, or all at once.
We must take it one piece at a time, and if
we choose our pieces wisely, the picture
can be visualized sooner.
Highlights of the Authors'
Recommendations
Within each chapter, authors provide their
interpretations of the most important re-
search directions in their fields of expertise.
Their recommendations can be placed
within the context of the various compo-
nents of risk of injury, and the distillation
of their suggestions are represented in ta-
bles 2-5. However, any comprehensive re-
search agenda will be severely limited by
resource constraints; major choices will
have to be made. Before outlining a strat-
egy for action, we must first outline our
. . . . . .
criteria tor pnor~t~zat~on.
Criteria for Prioritization
To achieve the primary objective of im-
proving our ability to quantify risks, mul-
tiple criteria were used to guide our
selection of future research efforts. Consid-
eration must be given to the relative impor-
tance of the information sought and the
feasibility of obtaining it. Importance may
be related to the seriousness of the health
risk or to the likelihood that it may result
from exposures near ambient levels. Im-
portance is also linked to societal needs.
Before industry utilizes a new technology
OCR for page 28
28
A Strategy for Quantifying Risk
Table 2. Chemical and Physical Properties
Topic
Formation and Transformation of Unregulated Pollutants
Vehicle exhaust
· Formaldehyde: There is a need for real-time concentration data under various
driving conditions in methanol-fueled vehicles operated with and without catalysts.
· Nitro-polycyclic aromatic hydrocarbons (nitro-PAHs): Kinetics of formation in the
exhaust system and dilution tunnel for diesel emissions should be determined.
· Diesel exhaust: Detailed characterization of particulate and gas-phase hydrocarbons
is needed.
· Diesel particulates: Research on particulate control technology is needed. Data on
particle size distribution, metal species, adsorbed hydrocarbons; and effects of
additives should be collected.
Atmospheric reactions
· Oxides of nitrogen: Further investigation of the transformations of oxides of
nitrogen under atmospheric conditions is needed. This topic is important for indoor
.
Author(s)
Johnson
Atkinson
environments as well.
PAHs and nitro-PAHs: Atmospheric transformation products of PAHs in gaseous
as well as adsorbed phases require study. Quantitative information on reaction
pathways leading to nitro-PAHs and on removal processes for nitro-PAHs is
needed.
· Aromatic hydrocarbons: The products arising from hydroxyl radical-initiated
reactions of the aromatic hydrocarbons should be identified.
Aerosol Processes
Data are needed on aerosol formation, particle size distribution, chemical compost- Graedel; Russell
lion, and chemical transformations.
Instrumentation and Analytical Methods
Real-time measurement methods should be refined in order to more accurately
quantify emission constituents.
Analytical techniques for the nondestructive, nonintrusive, in situ study of
transformation products of gaseous as well as particle-associated chemical species
should be improved.
or government agencies formulate regula
tory policies, certain scientific knowledge
may be desirable. Finally, some pieces of
information must be obtained before other
ideas can be pursued. For example, the
application of dosimetry models requires
that crucial input data be representative.
Some research topics are extremely im
portant, but unfortunately, with current
knowledge or methodology, their near
term solutions are not very feasible. For
example, it is essential that the science of
toxicology evolve to the point where com
plex mixtures can be analyzed for their
toxicity. A prerequisite will be the devel
opment of quantitative markers associated
with the onset of disease. If sufficient
knowledge and adequate methodologies
are available, then long-range research
goals should be pursued.
Johnson
Atkinson
A Proposed Research Strategy
If we cross-stitch the recommendations from
the various authors with the criteria for prior-
itization, we can begin to chart a course for
future research. Two major themes emerge
that appear to be at the heart of our objective
of generating an adequate scientific data base
for risk assessment: (1) an increased emphasis
on the quantification of exposure and dose,
and (2) an improved understanding of basic
disease processes.
Quantification of Exposure and Dose.
The development of molecular dosimeters
should constitute a long-range research ob-
jective. Currently, the formation of ad-
ducts has been exploited the most for our
purposes; however, validation of specific
adducts and identification of different do
OCR for page 29
Richard R. Bates and Ann Y. Watson
29
Table 3. Exposure Characterization
T.
OplC
Author(s)
Air Quality
Atmospheric evaluation
· The chemistry of atmospheric aerosol particles should be monitored in Graedel
more detail. Chemical differences as a function of particle size should be
determined.
· Routine monitoring of alcohol and aldehyde levels should be done in areas
where alcohol-based fuels are or will be in heavy use.
· Better mathematical descriptions of appropriate chemical reactions for use
in models should be made.
· Improved descriptions of pollutant transport and dispersion are needed for
street canyons and other complex situations in which air movement is
restricted. Inclusion of chemical reactions within a street canyon model is
also needed.
· The potential exposure of passengers in closed and open vehicles along Samson
roadways should be examined. Research should continue to focus on the
role of vehicular turbulence in initial dispersion of exhaust.
Indoor evaluation
· Indoor pollutants should be measured and their sources identified.
· Models should be improved for apportioning specific emission sources
to individual exposures and for relating outdoor to indoor air quality.
Exposure Assessment
Studies should be undertaken to provide information on the spatial and
temporal distributions of human populations as they relate to exposure. To
obtain this information, valid and reliable questionnaires should be developed.
Instrumentation
· Development and use of more suitable instruments for indoor and personal
monitoring are needed.
· A reliable, sensitive formaldehyde or aldehyde monitor should be developed.
Dosimeters
· Sensitive methods to detect adducts should be developed.
· Available biological measurement techniques should be adapted to air
pollution monitoring. The relationship among exposure, dose, and health
outcome requires better understanding.
Model Validation
Detailed and accurate sets of input data should be used to assess the adequacy
of current air quality and exposure models.
Graedel; Marnett
Russell
Russell; Samson
Graedel; Sexton and
Ryan
Russell; Sexton and
Ryan
Bresnitz and Rest;
Sexton and Ryan
Bresnitz and Rest;
Graedel; Sexton and Ryan
Graedel
Hecht; Kaufman; Marnett
Sexton and Ryan
Russell; Sexton and Ryan
simeters associated with other disease pro-
cesses are needed. To accomplish these
goals, more knowledge about the molecu-
lar events related to the onset of disease
should be obtained. Furthermore, if molec-
ular dosimeters are to serve as an effective
link between pollutant exposure and subse-
quent health effects, the quantitative rela-
tionship among exposure, dosimeter levels,
and effects will have to be established.
Therefore, until we are able to identify
and validate molecular dosimeters, more
emphasis should be placed on refining sur-
rogate measures of dose. For this, better
characterizations of exposure, as well as
improved models for dose assessment are
needed. It is becoming increasingly appar-
ent, however, that exposure estimation is
not as straightforward as initially supposed
and may have uncertainties as great as, or
greater than, those associated with dose/re-
sponse estimates. We need more informa-
tion about environmental levels of most
unregulated pollutants, indoor exposure
levels, and how to integrate indoor and
outdoor levels into a comprehensive pic-
ture of human exposure. This information
is also needed to conduct and interpret
experimental studies as well as to validate
dosimetry models.
OCR for page 30
30
A Strategy for Quantifying Risk
Table 4. Dose Assessment
Topic
Author(s)
Expanded Data Bases
Comparative analyses in normal species
· Morphometric measurements at all levels of the respiratory tract Schlesinger
should be done in adult humans and laboratory animals.
Comparisons among and within species should be made and
statistical variability determined.
· More emphasis should be placed on the upper respiratory tract,
including the development of dosimetry models for this region.
· Description of the liquid lining of the lung in humans and laboratory
animals is needed.
· Data on deposition and clearance kinetics of particles in different
species should be gathered.
· Scaling factors should be developed based on measurements of total
uptake of eases in different animal species.
Overton and Miller; Schlesinger
Overton and Miller; Schlesinger
Schlesinger
Ultman
--I- - <~
High-risk groups
· Morphometric measurements are needed for sensitive groups such
as the young, old, and diseased. New anatomic models should be
developed.
· Airflow patterns and distribution could be better described in Schlesinger
sensitive subgroups in laboratory animals and humans.
Diesel Exhaust and Particle-Associated Organics
Chronic studies should be undertaken at low concentrations of diesel
exhaust, and long-term clearance, translocation, and retention of
diesel particulates should be assessed. Coexposures to other pollutants
should also be conducted. Healthy animals and models of sensitive
populations should be evaluated.
The effects of carrier particles on the ultimate disposition of adsorbed
organics should be determined.
Rates of desorption of adsorbed compounds from inhaled particles
should be quantified.
Gases
Chemical reactions of specific pollutants with mucus, blood, and tissues
should be quantified.
Analysis of mass transport through individual diffusion barriers, Ultman
particularly the mucous layer, the bronchial wall, and the alveolar
capillary network is needed.
Noninvasive techniques to evaluate the transport of soluble and reactant Ultman
pollutants should be developed. These techniques coupled with
appropriate mathematical models can be used to obtain information
on regional inhomogeneities in dose and uptake.
Overton and Miller; Schlesinger
Schlesinger
Schlesinger; Sun, Bond, and Dahl
Sun, Bond, and Dahl
Overton and Miller; Ultman
Methods should be developed and experimental data should be obtained Bromberg; Overton and Miller
for dosimetry model validation.
The methodology to obtain useful ex-
posure information on the regulated pol-
lutants is now available. In contrast, en-
vironmental levels of most unregulated
pollutants have not been adequately char-
acterized. Because government standards
do not exist, there appears to be little
incentive to develop monitoring equipment
or programs. But we cannot be lulled into
inaction by thinking that trace amounts of a
particular substance will be of little conse
quence. For example, in the atmosphere,
the concentration of the hydroxyl radical is
only 10-6 to 10-7 ppm; however, the hy-
droxyl radical is one of the most reactive
atmospheric chemical species known and
participates in the majority of atmospheric
scavenging reactions. Biologists will have
to alert chemists and engineers to poten-
tially toxic pollutants that should be mea-
sured.
Data on the concentrations of regulated
OCR for page 31
Richard R. Bates and Ann Y. Watson
31
Table 5. Biological Responses
Topic
Author(s)
Noncarcinogenic Pulmonary Effects
Chronic diseases
· An epidemiologic cohort study of the effects of long-term exposure to oxidants on
respiratory morbidity should be conducted.
· Role of injury:
The pathophysiolozical significance of alterations in bronchoalveolar epithelial Last
Bresnitz and Rest
A ~ ~ _ _
permeability in animals and humans should be determined.
Mediators derived from damaged epithelial cells and inflammatory cells should be Last; Wright
identified and characterized by use of cell culture systems.
The relationships among the various stages of injury and changes in populations of Last; Wright
pulmonary macrophages and interstitial cells should be established.
· Regulation of repair:
The reversibility of abnormal collagen structure or deposition should be determined. Last
The relationship among digestive enzymes, connective tissue molecules, and abnor- Wright
mat lung tissue should be examined.
A better understanding of disease progression is needed. Lung structure and Last; Wright
biochemistry should be evaluated during a postexposure recovery period.
Asthma
· Panels of individuals with bronchial hyperreactivity should be studied to determine
whether exposure to ambient oxidants affects respiratory symptoms or morbidity.
· Controlled clinical studies should be conducted with asthmatics exposed to chamber Bromberg
atmospheres similar in composition to ambient atmospheres associated with
increased symptomology.
Bresnitz and Rest
· The effect ot the presence of allergens either during or after pollutant exposure Bromberg
should be determined in extrinsic asthmatics.
· The role of airway C-fiber sensory systems in ozone effects in epithelial properties, Bromberg
bronchial reactivity, and airways inflammation should be clarified.
Respiratory infection
· Epidemiologic surveys should be performed in high-risk populations using serologic
and cultural tests to confirm infection.
· In animal or human studies under controlled laboratory conditions, components of
respiratory antiviral defense mechanisms should be evaluated.
Carcinogenic Effects
~. . . . .
Bresnitz and Rest;
Pennington
Pennington
Critical data on the corrections used in extrapolations should be obtained for Kaufman
.. . . . ... . . . ~
. .
quantitative assessments.
Diesel emissions
· Additional studies should be performed on the carcinogenicity of diesel exhaust.
· Methods should be developed for assessing the carcinogenicity of mixtures.
PAHs and nitro-PAHs
· Structures of the major DNA adducts and protein adducts formed from
representative PAHs and nitro-PAHs should be identified.
· Under conditions of chronic administration of PAHs or nitro-PAHs to experimental
animals, the relationship between DNA or protein adduct formation and tumor
development should be determined.
· In order to determine the feasibility of monitoring DNA and protein adducts in
humans, pilot studies in individuals potentially exposed to PAHs or nitro-PAHs
should be performed.
· Major pathways of metabolic activation and detoxification should be determined.
Aldehydes
· A chronic inhalation toxicology study of acrolein should be undertaken in rats,
with emphasis on carcinogenicity.
· A chronic inhalation toxicology study of mixtures of formaldehyde and acrolein
should be undertaken in rats and hamsters, with emphasis on carcinogencity.
· Experiments should be undertaken in cells cultured from the upper respiratory tract
to determine the mechanisms by which aldehydes exert pathological changes.
Kaufman
Hecht
Marnett
(Table continued next page.)
OCR for page 32
32
A Strategy for Quantifying Risk
Table 5. Continued
Topic
Author(s)
Other
Coronary heart disease
· Using ongoing cohort studies or other existing data bases, increased risk of heart
disease should be evaluated in persons exposed to varying levels of automotive
emissions.
· The effect of automotive emissions on various components of atherogenesis should
be evaluated in animal studies that use cynomolgus monkeys.
Neurobehavioral effects
· Tests to detect functional disorders and sensory impairments should be applied in a
systematic manner to evaluate the magnitude and prevalence of behavioral alterations
due to components of automotive emissions.
Clarkson
Wood
as well as unregulated pollutants are inad-
equate for indoor environments. Although
residences and office buildings are the most
logical indoor environments to investigate,
the interiors of vehicles should also be
considered. Some segments of the popula-
tion (for example, commuters and truck
drivers) spend considerable time, on a reg-
ular basis, in vehicles where concentrations
of pollutants could be high.
In addition to improved charactenzat~on
of human microenvironments, more infor-
mation on the amount of time spent in
various locations is needed. Documenta-
tion of this sort is inadequate for different
sectors of the population, but it is not
technically difficult to obtain. Individual
activity constitutes an essential component
in the assessments of human exposure; the
construction worker experiences a mark-
edly different exposure than the business-
person who walks down the same street on
his or her way to lunch.
To improve dose assessment, descriptive
data bases about respiratory tract anatomy
and ventilation must be expanded. Pollu-
tant dose to the lung depends on airway
geometry. Most particle deposition and gas
transport dosimetry models use the sym-
metrical lung model developed by Weibel
in 1963 (Weibel 1963~; this model is based
on measurements from one human speci-
men. Recently, more elaborate models
have been proposed that attempt to account
for variations in lung geometry. Better
anatomic models will improve the accuracy
of dose estimates. The development and
validation of more realistic models requires
representative input data, which are largely
unavailable. In addition, measurements of,
and a greater emphasis on, the upper air-
ways should be pursued. Large particles
and their associated organics, as well as
soluble gases, will contact respiratory tis-
sues, and perhaps gain access to the circu-
lation, in the upper airways. Without accu-
rate physical and biochemical descriptions
of the liquid lining, the fluid's role as a
protective layer for underlying epithelial
tissues cannot be determined. Finally, the
distribution of inhaled particles and gases is
also influenced, but poorly characterized,
by ventilation. The effects of increased ven-
tilation have a practical significance for
physically active individuals.
Measurements taken in a systematic
manner on more specimens within and
between species are needed. Experimental
data from animals cannot be confidently
extrapolated to humans until intra- and
interspecies variability has been deter-
mined. Data are needed on normal individ-
uals as well as on high-risk subjects. Few
descriptive data are available on the young,
elderly, or special disease groups. The an-
atomic and physiological status of these
individuals may profoundly influence pol-
lutant dose to their respiratory systems.
Basic Disease Processes. The contribu-
tion of any pollutant to the initiation or
exacerbation of a disease cannot be accu-
rately determined until we have more in-
sight into the pathogenesis of that disease.
By focusing our efforts on early events, we
would like to identify the cellular and mo
OCR for page 33
Richard R. Bates and Ann Y. Watson
33
lecular alterations that may link pollutant
exposure to ill health. This information
would also aid in the identification of sus-
ceptible subpopulations. Thus, real prog-
ress in defining human risks is not likely to
occur until the cellular and biochemical
mechanisms of various diseases are better
characterized.
Early events of many respiratory diseases
involve damage to epithelial tissues; altered
epithelial permeability is implicated in the
etiology of an asthmatic attack; damaged
epithelial cells release biochemical signals
for inflammatory cells, which in turn ap-
pear to regulate interstitial cell function;
and proliferation of damaged cells is asso-
ciated with cancers of the respiratory tract.
Therefore, it is imperative to focus more
attention on the mechanisms by which the
epithelium translates insult to disease. Once
the integrity of the pulmonary epithelium
is compromised by injury, underlying tis-
sues are more susceptible to harm from
inhaled substances. In addition to their func-
tion as a protective barrier, the extent to
which epithelial cells actively contribute to
pathological processes is unclear. The recent
evidence that interactions between epithelial
and interstitial tissue layers affect mutual
structure and function suggests a more active
role for epithelial cells. In addition to the
identification of biochemical mediators in
response to injury, it is also unknown how
much of the specificity of outcome resides in
the signals generated. Whether a pollutant
induces the formation of free radicals or
macromolecular adducts may influence tissue
responses normal and abnormal.
Another important area for investigation
is improved characterization of the inflam-
matory response. One cellular component,
the macrophage, produces growth factors
and mediators of activity for several other
cell types. These macrophage-derived
products probably play a role in the devel-
opment of fibrosis, emphysema, small air-
ways disease, and atherosclerosis. In addi-
tion, evidence suggests that inflammation
influences bronchial hyperreactivity. The
balance between defense and dysfunction is
more than likely regulated, in part, by
biochemical signals generated during an
inflammatory reaction. In addition to the
characterization of these signals, the cir-
cumstances that govern their production
require clarification. For example, the tim-
ing of insults during repeated exposure
may be critical for normal repair. Ideally,
we would like to elucidate the conditions
that solicit physiological "backfire"-that
is, when defense mechanisms no longer
operate to protect the organism, but in-
stead, promote disease. Inflammatory
events after exposure to initiating agents
should be examined in established model
systems of specific disease states.
Finally, research efforts should continue
in the arena of genetic control mechanisms.
The explosion of research in this area pro-
vides toxicologists with a potentially im-
portant methodology-the use of DNA
adducts as dosimeters of carcinogenic inter-
actions with target sites. However, we
must still face the mystery and chal-
lenge-of relating genome alterations to
human disease. It is not enough to just
identify adducts relevant to emissions or
emission products. Genetic factors that
affect individual susceptibility should be
explored further. Tremendous variation
exists among human individuals in the
metabolic activation of substances to carci-
nogenic compounds, the molecular basis of
which has not been characterized. In addi-
tion, factors that influence the capacity of
pulmonary tissues to repair genome dam-
age should be investigated. Comparisons of
biochemical profiles between sensitive and
resistant strains of laboratory animals may
aid in the elucidation of these factors. In
conjunction with animal studies, increased
emphasis should be placed on applying
current methodologies to or developing
more appropriate analytical techniques for
human samples. Finally, if we hope to
assess the carcinogenic potential of com-
plex mixtures of airborne pollutants, we
need a better understanding of the molec-
ular basis for the actions of promoters and
. . q
cocarc~nogen~c su Stances.
Among these interconnected areas of re-
search lie clues to many fundamental prob-
lems in abnormal tissue structure and func-
tion that constitute toxicity and disease.
Opportunities for more quantitative toxi-
cology should be pursued using the rapid
. . ~
OCR for page 34
34
A Strategy for Quantifying Risk
advances in scientific methodology and the
recent knowledge of molecular and cell
biology.
Summary
Currently, our abilities to obtain data for
risk determination directly from humans is
limited. Epidemiologic studies that evalu-
ate chronic diseases should be limited until,
or unless, adequate estimates of exposure and
early biological markers are available. Be-
cause of the possibility of correlating expo-
sure with outcome, the incidence of asth-
matic attacks or respiratory infection holds
the most promise for epidemiologic investi-
gation, assuming that satisfactory methodol-
ogy for the detection of infection can be
applied. Controlled clinical trials are best
applied to the analysis of neurobehavioral
effects or bronchial hyperreactivity. Even in
cases of asthma, data collection is restricted to
a subpopulation of mildly affected subjects.
We need more information on what deter-
mines the severity of the disease.
The example of asthma illustrates an-
other important point; results in clinical
studies have not supported epidemiological
evidence of an effect of oxidants on airway
sensitivity. Whether we have been unable
to accurately duplicate ambient atmo-
spheres\ or identify appropriate susceptible
individuals for study is not known. Both
factors require additional consideration
when conducting further studies.
For most health effects animal expen-
ments must be used as a source of informa-
tion from which human risks can be ap-
proximated. But how can the data be
linked more effectively to the human situ-
ation? Since risk is related both to dose and
to sensitivity, each should be considered a
research area that might improve our abil-
ity to quantify human risks from animal
experiments. Fruitful directions that would
improve dosage extrapolation are readily
apparent. They lie partly in further devel-
opment and application of mathematical
models for comparing pulmonary deposi-
tion in laboratory animals and humans. As
the harmful components of automotive
emissions are identified, biological markers
of their presence in tissues can also be
developed and used to compare target site
levels. Definitive comparisons of sensitiv-
ity of laboratory animals and humans to
equivalent target site doses probably awaits
better understanding of the molecular de-
terminants in chronic disease development.
Elucidation of missing links between acute
effects and chronic disease would provide
investigators with a powerful tool early
indicators of subsequent injury. Knowing
biological markers and factors that deter-
mine individual sensitivity will not only
provide better extrapolations from animals
to humans, but will also improve the fea-
sibility of using epidemiologic approaches.
Ideally, a comprehensive research strat-
egy would link molecular dosimetry to
disease in laboratory animals and finally to
exposure in humans. This strategy is best
illustrated using PAHs carcinogenic com-
ponents in vehicle emissions. First, it will
be necessary to identify the appropriate
DNA and protein adducts in PAH-exposed
animals. Second, and probably most cru-
cial, will be to determine the quantitative
relationship between adduct levels and tu-
mor formation. Finally, if a predictive cor-
relation can be made between marker and
disease, it will then be appropriate to apply
these methods to PAH-exposed humans.
Conclusion
Each chapter in this book addresses some
aspect of the problem of assessing exposure
and risks to humans from automotive
emissions and their transformation prod-
ucts. Invariably, and inevitably, the authors
found that the research needed to answer
questions about automotive emissions IS
also applicable to a much broader array of
issues about human exposure and health.
The development of new techniques and a
better understanding of atmospheric chem-
istry and physics, all of which would im-
prove our knowledge of exposure to auto-
motive emissions, are also applicable to
other sources of air pollution. Similarly,
the methods and the knowledge needed to
assess health risks from motor vehicle
emissions are also useful to evaluate the
OCR for page 35
Richard R. Bates and Ann Y. Watson
35
effects of other substances in the environ-
ment. Thus, rather than there being a sharp
boundary of research questions around au
. . . . , .
tomot1ve emissions, a series ot concentric
circles that overlap other topics of environ-
mental exposure and health exists.
Before proceeding too far toward the
more peripheral circles, it is worth return-
ing to the basic question: How important
are automotive emissions as a risk to hu-
man health? Referring to air pollution in
general, a committee of the National Re-
search Council recently concluded that
"evidence from controlled human expo-
sures, toxicology, and epidemiology is suf-
ficient to warrant concern tliat current air
pollution still produces substantial adverse
health effects in some segments of the U. S.
population." Furthermore, "The Commit-
tee finds that current air pollution can cause
acute and perhaps chronic health effects,
particularly respiratory effects, in the pop-
ulation of the United States. Respiratory
disease is a major cause of work loss and
disability. Even if only a small proportion
of very prevalent disease is due to air
pollution, the absolute amount of illness
that could be prevented by reducing air
pollution would be large" (National Re-
search Council 1985~. These comments
were balanced by another statement: "The
impact of ambient air pollutants on the
total respiratory disease burden in the
United States must be small relative to the
impact of cigarette smoking, and occupa-
tional exposures might also have greater
effects than pollution of ambient air" (Na-
tional Research Council 1985~.
Motor vehicle emissions are responsible
for a substantial proportion of atmospheric
exposure to carbon monoxide, nitrogen
oxides, volatile organic compounds, and,
in urban areas, total suspended particulate
matter. Atmospheric organic compounds
and nitrogen oxides derived from automo-
tive emissions also contribute significantly
to the formation of ozone. Thus, to the
extent that these substances are a source of
the health concerns of the National Re
Correspondence should be addressed to Ann Y. Wat-
son, Health Effects Institute, 215 First Street, Cam-
bridge, MA 02142.
search Council Committee' automotive
emissions must also be of concern. Ques-
tions about this issue have been raised, but
not fully answered, as the chapters of this
book indicate.
Much less information is available on the
hundreds, or perhaps thousands, of indi-
vidual organic chemicals that contribute to
the volatile organic and particulate fractions
of atmospheric pollution. Neither the ex-
posure levels, the atmospheric reactions,
nor the health effects of many of these have
been well characterized. Among these ma-
terials are the PAHs, nitro-PAHs, and al-
dehydes which are discussed in this book,
as well as many other chemicals. It seems
likely that many of the organic chemicals
derived from motor vehicle operation are
present at such low atmospheric levels that
they do not threaten human health, even in
sensitive individuals. On the other hand,
highly toxic chemicals can harm health
even at very low concentrations, especially
when they are mixed with other substances
and synergism occurs. Some balance needs
to be struck between unnecessary Research
on trivial problems and insufficient research
on possibly important ones. Locating the
balance point is not easy. It depends on a
continuing dialogue between the experts on
atmospheric exposure and the experts on
biological effects of toxic substances. The
former can contribute information about
what is in the atmosphere, the latter on
what may be important for health. Thus,
each group can help guide the research
priorities of the other. We hope that this
book will contribute to the dialogue.
References
Klaassen, C. D., and Doull, J. 1980. Evaluation of
safety: Toxicologic evaluation, In: Casarett and
Doull~s Toxicology: The Basic Science of Poisons a
Doull, C. D. Klaassen, and M. O. Amdur, eds.),
2nd ea., pp. 11-27, MacMillan, New York.
National Research Council (Committee on the Insti-
tutional Means for Assessment of Risks to Public
Health). 1983. Risk Assessment in the Federal Govern-
ment: Managing the Process, p. 191, National Acad-
emy Press, Washington, D.C.
National Research Council (Committee on the Epi-
demiology of Air Pollutants). 1985. Epidemiology
OCR for page 36
36
A Strategy for Quantifying Risk
and Air Pollution, p. 224, National Academy Press,
Washington, D.C.
Paracelsus. 1958. Selected Writings U Jacobi, ea., and
N. Guterman, trans.), Bollingen series XXVIII, p.
290, Pantheon Books, New York.
U.S. Environmental Protection Agency. 1984. Na-
tional Emissions Report, 1981, National Emissions
Data System of the Aeromatic and Emissions Re
porting System, Report EPA-450/4-83-022, Re-
search Triangle Park, N.C.
U.S. Environmental Protection Agency. 1986. Na-
tional Air Quality and Emissions Trends Report,
1984, Report EPA-450/4-8~029, Research Triangle
Park, N.C.
Weibel, E. 1963. Morphometry of the Human Lung,
Springer-Verlag, Berlin.
Representative terms from entire chapter:
respiratory tract
